Interactive startup on F10. Pressing "I" doesn't work

Alan Evans ame.fedora at
Fri Apr 3 14:26:09 UTC 2009

Clark Martin wrote:
> Alan Evans wrote:
>> I would have thought that the key-repeat wouldn't differ functionally
>> from rapidly tapping the key.
> Most computer keyboards transmit to the computer a key down and key up for
> each key.  This is true for modifiers as well (shift, control, etc).  The
> computer decides if you've held the key down long enough to repeat and also
> determines how fast to repeat the key.  Which means if the keyboard buffer
> is cleared AFTER you've pressed the key the program won't see the key down
> and therefore won't consider a key as repeating.
> In normal use when the buffer is never cleared typing fast and key repeat
> would be equivalent.

Are you suggesting that the init script clears the keyboard buffer at
the driver level? How would it even do that?

If I write an application that's reading stdin, that app is certainly
not worrying about key press events and key release events to
determine if modifiers are at play. It's just taking chars out of

As I understand it, if that application "clears the buffer" then it's
doing something along the lines of calling flush(). This doesn't cause
the system-level keyboard driver to forget that a key has been
pressed. That driver continues injecting chars into stdin even after
my app has flushed the buffer.

I would think that it would be considerably harder to detect a
modifier press or release in a shell script.

Maybe I'm completely wrong about all that...

More information about the users mailing list